In [12]:
label_registry = get_label_registry()
df = pd.read_csv(path)
df = df[df['cellid'] != 9]
df['esoh_c3_Cn'][df['esoh_c3_Cn'] > 3] = np.nan
df['esoh_c3_CnCp'][df['esoh_c3_CnCp'] > 1.2] = np.nan
df['var_q_1c_c100_c10_mah'] = df['var_q_1c_c100_c10_ah'] * 1000
df_rt = df[df['is_room_temp_aging'] == 1]
df_ht = df[df['is_room_temp_aging'] == 0]
xvar_array = ['form_last_charge_voltage_after_1s',
'form_final_discharge_capacity_ah',
'form_first_charge_capacity_ah',
'form_first_cycle_efficiency',
'form_coulombic_efficiency',
'form_qc_minus_qd_ah',
'form_6hr_rest_delta_voltage_v',
'dcr_10s_5_soc_at_c3',
'dcr_10s_10_soc_at_c3',
'dcr_10s_90_soc_at_c3',
'dcr_10s_90_soc_at_c3',
'dcr_10s_5_soc_at_c3_chg',
'dcr_10s_0_soc_at_c3_chg',
'dcr_10s_0_soc_at_c3',
'var_q_1c_c100_c10_mah',
'rpt_c3_delta_v']
# cyc_idx_rpt_arr = [3, 56, 159, 262, 365]
# [xvar_array.append(f'var_q_c20_c{x}_c3_ah') for x in cyc_idx_rpt_arr[1::] ]
# [xvar_array.append(f'esoh_c{x}_Cn') for x in cyc_idx_rpt_arr ]
# [xvar_array.append(f'esoh_c{x}_CnCp') for x in cyc_idx_rpt_arr ]
# [xvar_array.append(f'esoh_c{x}_x0') for x in cyc_idx_rpt_arr ]
# [xvar_array.append(f'esoh_c{x}_x100') for x in cyc_idx_rpt_arr ]
# [xvar_array.append(f'esoh_c{x}_y0') for x in cyc_idx_rpt_arr ]
# [xvar_array.append(f'esoh_c{x}_y100') for x in cyc_idx_rpt_arr ]
# [xvar_array.append(f'esoh_c{x}_neg_excess') for x in cyc_idx_rpt_arr ]
# Also try: 'thickness_mm'
# yvar = 'form_qc_minus_qd_ah'
yvar = 'cycles_to_70_pct'
df_array = [df_rt, df_ht]
color_fast_array = [COLOR_FAST_RT, COLOR_FAST_HT]
# Loop over temperatures
for df, color_fast in zip(df_array, color_fast_array):
# Loop over x-variables
for xvar in xvar_array:
model = LinearRegression()
x = df[xvar].values
y = df[yvar].values
idx_to_keep = ~np.isnan(x)
x = x[idx_to_keep]
y = y[idx_to_keep]
corr, _ = pearsonr(x, y)
model.fit(x.reshape(-1, 1), y)
x_fit = np.linspace(np.min(x), np.max(x), 100)
y_fit = model.predict(x_fit.reshape(-1, 1))
idx_base = np.where(df['is_baseline_formation'] == True)[0]
idx_fast = np.where(df['is_baseline_formation'] == False)[0]
plt.figure(figsize=(4, 4))
plt.plot(df[xvar].iloc[idx_fast], df[yvar].iloc[idx_fast],
linestyle='None', marker='o', markersize=7,
color=color_fast, label='')
plt.plot(df[xvar].iloc[idx_base], df[yvar].iloc[idx_base],
linestyle='None', marker='o', markersize=7,
color=COLOR_BASE, label='')
if corr**2 > 0.5:
plt.plot(x_fit, y_fit, color=COLOR_REF, label='')
plt.plot(np.nan, np.nan, label=rf'$\rho$ = {corr:.2f}',
marker='None', linestyle='None')
leg = plt.legend(frameon=False,
handletextpad=0, handlelength=0, fontsize=15)
# plt.xlim(label_registry[xvar][1])
plt.xlim((None, None))
plt.xlabel(label_registry[xvar][0])
plt.ylabel(label_registry[yvar][0])
